Notes on: Grills, C., Longshore, D (1996). Africentrism: Psychometric Analyses of a Self – Report Measure. Journal of Black Psychology 22 (1): 86 – 106

Dave Harris

[This is where I first encountered Nguzo Saba as a definition of African values. V. disappointing! Lots on Google about them. These people tried to operationalise the terms into a psychometric test and then see how important they were to Black drug-users!]

Social services are more effective if they can be targeted according to relevant cultural characteristics, as a number of earlier reports indicate, for example on cultural competence in substance use prevention. This requires tools to measure these characteristics, together with individual variability. This attempted to develop a self-report measure of Africentrism, 'the degree to which a person adheres to the Nguzo Saba (seven principles) in African and African-American culture' (87).

They define culture as elements of a group's history, values, knowledge and norms that have become meaningful explicitly or explicitly, and see it as a constant at the group level, although variable at the individual level. This variability means evaluation is difficult and leaves only group level analysis with individual differences 'as error variance' (88).

'Certain principles in African and African-American culture have been termed Africentric' [a number of references] and as a suitable basis to develop services [more references] their working definition is based on the Nguzo Saba:

Umoja (unity), unity in the family community nation and race
Kujichagulia (self-determination ), self-determination, wanting to speak for yourself instead of being defined by others
Ujima, (collective work and responsibility) building community and taking collective responsibility for people's problems.
Ujamaa (cooperation in econ activity), cooperative economics to build shops and other businesses and profit from them together
Nia (purpose) , to build and develop the community and 'restore our people to their traditional greatness' as a collective vocation.
Kuumba (creativity), to leave the community more beautiful and beneficial than when we inherited it

Imani (faith), to believe 'with all our heart' in our people parents teachers and the 'righteousness and victory of our struggle'.

The principles are apparently grounded in'Kawaida Theory' which is 'a synthesis of tradition and reason in nationalist, pan- Africanist and socialist thought'. They are seen as the minimum set of values needed to build an Africentric family community and culture and are to be taken as codes of conduct for every day living. They might be manifested through, for example, voluntary community service (ujima). We need to further turn these into tools to help assessment and targeting [and at this moment they turn to talking of Africentrism -- why?]. There are already certain measures of ethnic identity and values but these might not be adequate, not specific enough, in one case to African-American or Africentric values. They might also be mixed in terms of their psychometric properties. The link between racial identity and Africentrism is unclear and they may not correspond at all (89). Some measures are better than others at differentiating between Eurocentric and Africentric thought. Others have been tested mostly with college students requiring high levels of literacy. They wanted something usable and brief and easy to administer.

The principles served as a guideline for item development, together with various 'African-centred cultural precepts' derived from the literature. This led to 25 items (Likert-type ) with options offering ranges of agreement. They 'worded some items positively so that agreement indicated views consistent with Africentrism' and others negatively [all pretty leading though?] . They then constructed two forms depending on whether they were working with specifically African-Americans. Examples of operationalisms include 'my families needs are more important to me than my own needs', 'I'm more concerned with my own goals and with helping other people reach there is' 'I'm doing a lot to improve my neighbourhood', 'I have more confidence in wide professionals like doctors and teachers than in [group] professionals' [a lot of them just turn on the difference between stressing your own goals and those of the community or the group -- hence the unsurprising underlying Factor 1 below]. In more detail:


1. My families needs are more important to me than my own needs
2a a [group] should make the community better than it was when they found it
 2b people should make the world better than it was when they found it
 3 All the problems of other [group] are their problems not mine…
 5b I am more concerned with my own goals than with helping other people reach theirs…
X(?)  I have more confidence in White professionals, like doctors and teachers than in [group] professionals…
13 I must do all I can to restore [group] to their position of respect in the world…
15b a it hurts me when I see a person like me discriminated against…
15b it hurts me when I see another [group] person discriminated against ...
17 B society would be better off if people just work on their own goals

The second form was meant to apply to more ethnic groups and to produce 'race neutral items applicable to both African-Americans and non-African Americans' (92) 4 items were omitted (relating to African American groups] and others slightly reworded.

They then did Cronbach alpha measures of reliability, and opted for quite high measures, even if they were 'seldom met in measures of psychosocial attitudes' because they wanted to 'assess differences in status among individuals and detect the extent to which an individual status is changed'.

They assessed validity in various ways — 'known groups validity' involves comparing scores between groups already known to be different on the constructs, such as white people, with the assumption that African-Americans should get higher scores. Another group were studying African-American history and again this was expected to produce a higher score [they were also non-prison inmates?] . 'Construct validity' depends on the assumption that adherence to Africentric principles like spirituality and communalism is conceptually related to existing formulations of ethnic identity in which ethnicity is salient to one's sense of well-being, conduct, purpose in life, and feelings of connectedness to others' (93). This means we can get construct validity by correlating it with ethnic identity [I thought they said there was no strong correlation earlier?]. One scale explores the extent to which people are committed to their ethnic identity, how involved in traditions and activities they were, and whether they felt affirmative toward the group. A particular measure  [MEIM] was chosen for its 'reliability and validity among... African-Americans '. They then selected five items [presumably on the basis that these were most highly correlated with total scores, and noted two items that produced consistently low correlations with African-Americans.

They carried out four studies, three in a pilot exercise to assess reliability and ease of administration and to explore the validity. The fourth one was the real thing based on a larger sample.

Study one was on African-Americans, 29 in a sample involved in 'substance use recovery group in Los Angeles' (94).  In one pilot interview, they used current and former substance users and paid each one $10. They used dyadic interviews. One interview was Anglo, two were Asian and Latino, and the remaining five were African-American. If the respondent 'initially expressed indecision (e.g. don't know)' they were probed until they produced a codeable answer. Each form took about four minutes to go through, uncomfortable responses were rare and usually followed after mild probing. However the sample was too small.

The second pilot assessed reliability in another African-Americans sample and did the comparison with whites. This time there were 57 African-Americans in the sample and 21 whites, in a 'methadone maintenance clinic' and comments on the form. Here the alpha coefficient was .62 among African Americans, but items/index correlations were low for four items, although they decided to keep them. For whites, alpha was not .26 even after the deletion of dubious items. However, mean scores were similar for both groups [so what follows?].

The third study looked for reliability and known groups validity. 25 African-Americans self-administered the form, and they were studying African-American history and culture [in prison?]. This group produced a high alpha coefficient 0.82 and 'most items correlated well (R = 0.30 or higher) with the total score, although again there were some items that did not correlate well, and means for this group were higher than for the others. They interpret this to mean that 'Africentrism scores will be higher among African-Americans, for Africentric principles are presumably more salient' (97).

So on they went to the final study with a larger sample size and with further testability in mind. They wanted to compare scores among African-Americans and Whites who 'were known to be similar in socio-economic background' and they also wanted to correlate between their measure and the MEIM.

Before we go any further, note the details of study:

How do the questions rate as operationalisations of the principles?  Some obvious problems -- eg defining the 'community' or group that is being referenced here -- all Black people? Other drug-users? Black people in LA? How could they be expected to know about pan-African issues? The questions are pretty leading, demanding yes no whereas most people would want to say -- it depends? Who would want to confess to being a selfish bastard?

Items four and 12 to 14 were omitted on Form B— that is referring to the unity of the African race [4] and questions about building and maintaining their own communities and restoring their group to positions of respect, or shopping at group businesses] Some attracted 'reverse scoring' — 3a, 3B, 5a, six, 8a, 8b, 11, 17a, 17b. [I am not sure what thismeans -- that they were definitley not rated as agreeable?]

The final real study compared scores among African-Americans and Whites of similar socio-economic background and then constructed validity tests by correlating with the MEIM. They chose 78 African-American arrestees and 93 white arrestees, mostly men in both cases, and did dyadic interviews with them. The sample had been in jail at the time. They had been interviewed as a result of another survey of drug use. They did not pay anybody.

Any differences between the scores 'might have been confounded by group differences in social class, reading level or other socio-economic characteristics' and they checked for these [they seem to have gathered data to give them 'group means' on things like income levels and highest grade achieved]. They think that the samples 'appear to be similar on the socio-economic characteristics available to us' (97). They checked reliability through internal consistency again and noted those scores where there were low correlations between item and total score. Alpha coefficients were pretty good for both groups, or at least 'acceptable' (98).

They then calculated scores for each group on the items that were common to both, and then compared means. They recognise that 'this comparison can be questioned' (98) because although the items are the same, meanings might 'diverge in subtle ways' and the presence of the extra four items on the first form could have altered the meanings common to both. However they 'believe the findings are instructive for exploratory purposes'. Mean scores were 2.91 among African-Americans, 2.79 among whites. However, the team argue that this is 'significantly higher' for the first group. In terms of construct validity African-Americans who scored high on Africentrism also scored 'significantly higher on ethnic identity achievement, ethnic behaviour and affirmation' [but significance levels seem a bit low, p> .001 in each case].

Then they did some factor analysis to see if 'variation in the items set is consistent with the underlying construct'. They had already noted that some items did not correlate highly with index (total) scores, so  deleted those this time. They pursued 'a principal components analysis' to extract initial factors and chose 4 with suitable eigenvalues. They say that this 'four factor solution might overestimate the number of meaningful factors' however, because the increase in significant eigenvalues 'is an increasing function of the number of items' (99). They therefore looked for a particular 'elbow' which might separate these factors, a 'point of discontinuity, below which the remaining factors begin to explain relatively little variance' and found that between the first two factors. Indeed, 'a one factor solution may be the most appropriate' although 'a multifactor solution is also defensible'

Then they did some lovely rotations to check which items loaded best onto the factors. Nine loaded best on the first factor, 4 on the second. They named this factor as 'an individualism – communalism dimension'. Overall this one factor solution 'is the most meaningful way to account for item relationships [and capture]… The underlying construct of Africentrism'. However it is 'defensible to suggest the relevance of another factor, 'individualism – collectivism' [doesn't seem very different to me]. They really need larger samples and better choices for items [maybe], so the results are 'provisional but instructive' (100).

 Generally they're quite pleased with their Alpha coefficients, even though none reached the 'recommended minimum when individuals are to be compared or changes to be assessed' (101), but future research might develop stronger ones. Overall they had 'mixed findings [and so] we cannot conclude that the Africentrism measure is reliable for any research purpose among Whites', but this is no problem really because they were mainly interested in Blacks.

Strong ethnic identities 'are widely viewed as important to a healthy sense of self' especially, 'involvement in group activities and traditions, and group attachment or pride'. So they should expect to find positive relations with the MEIM, and they did. They now think 'the correlations were high enough to suggest that adherence to Africentrism is an integral part of ethnic minority for many African-Americans', although they have found something new, that Africentrism is a separate aspect of ethnic identity conceptually and empirically.

They also found expected differences between the African-American groups, study one, and the students of study three. Study two produced particularly low alpha coefficients, although they still think that this is 'additional support for the validity of the Africentrism measure' because they see culture as a unified field, even though particular groups may only share certain beliefs and behaviours from within it — nevertheless, 'one culture remains distinguishable from others, and meaningful to its members, as a particular configuration of elements' (102). In this light, the lack of coherence in item scores of whites 'actually makes sense, and showed that the items 'did not represent a unified cognitive set for them, whereas it did for African-Americans'. However higher alphas in study 4 raises the possibility that the low alpha in study two 'was a fluke' and more research is needed.

They are happy that their forms seem reliable 'for the purpose of measuring differences in Africentrism between groups of African-Americans', were nice and easy, and they seem to survive basic tests of validity. They already have another form which is a revision of the earlier ones which they want to test even further. The results are encouraging enough to be suggested to other researchers, and to be used in formal education programs with those who use drugs or alcohol. However larger samples are needed to get better ('more persuasive') factor analyses. They think that the factors they've identified should be focused on in particular and there should be more matching by demographic traits.

Nevertheless, they chose a measure for use with a particular group and it was convenient. It might not be 'generalised to African-Americans outside this target group. In fact ethnic identity may be problematic among African-Americans who use illicit drugs or alcohol' (103) so broader tests in the general African-American population is required. They also need to further work on reliability over time, to do more exploration of the influence of the mode of administration and the ethnicity of the administrators — nothing was found in this study but they spent little time on debriefing. They would like to do some research among non-African-Americans as well [not just white people?] and compare the results with African-Americans. If their item scores are not consistent, for example, and alpha coefficients low compared to African-Americans this will be 'useful'. If the two groups compare this will enable further tests of intergroup validity.

They think that 'these findings suggest that Africentric values are an integral yet distinguishable aspect of ethnic identity for many African-Americans' (104) so a measure of it should be used in evaluating programs of intervention. They have two possibilities at the moment — that highest scorers will respond more readily to interventions that are 'African centred'; that low scorers may gain more from an African centred intervention [as well] because they might acquire 'a healthier ethnic identity through exposure to African centred material'. Both hypotheses should be tested with pre-and post change questionnaires.

[As usual, all the results support the initial views of the researchers, even  negative ones, there is enough evidence to prolong discussion and more research is needed].